Leaching Rate Prediction And Operating Performance Assessment For Hydrometallurgical Processes With Missing Data

With the continuous integration of informatization and industrialization,the processindustry,as the pillar industry of the national economy,is developing towards an efficient,green and integrated direction.However,the rapid growth of economy and the continuous promotion of industrialization lead to the shortage of nonferrous metal resources,and the problem of resources has become the fundamental problem of affecting China’s sustainable development strategy.How to make use of low-grade mineral resources economically and effectively to ensure the highest economic benefits under the normal operation of the process is the focus of current research.Assessment of operation performance is to answer how optimal the current operating performance is under normal operating conditions.When the process is operating at a non-optimal performance,non-optimal cause identification is carried out,which offers guidance to bring the process back to the optimal performance.In order to solve the problem of missing data in the production process of hydrometallurgical industry,this paper proposes a Leaching rate prediction method in cyanide leaching process with missing values and an operating performance assessment method among whole process of Hydrometallurgy based on robust Bayesian method.This paper proposes a combined prediction algorithm based on RegionBoost for missing values.Various imputation techniques are used to fill the historical dataset with missing values,and a series of single prediction models,as members of the combined prediction model,are built on the basis of filled dataset.In order to distinguish the samples with different degrees of missing,this paper also proposes a new method to calculate the degree of discrepancy between two samples.This method is used to obtain the region where the online sample is located,and the weight of each single prediction model is determined according to its prediction accuracy in this region.Therefore,it realize the self-adjustment of the weight of each single prediction model.In the process of operating performance assessment of the whole process of hydrometallurgy with missing data,the operating performance assessment process is divided into two stages:offline modeling and online assessment.First,the whole process of hydrometallurgical production is divided into three sub-blocks,for each sub-block,the Robust Bayesian Method is used to construct the assessment model.When conducting online assessment,the online data with missing value need to be pre-processed,and the RegionBoost method and the method of calculating discrepancy mentioned above are used to obtain the region where the online sample is located.Comparing the assessment results of the data in the region after filled by different method,using the best method to fill the online data and input the evaluation model,the result obtained is the assessment result of current online data.When the process is assessed at a non-optimal performance,non-optimal cause identification is carried out through calculating the relative deterioration degree which offers guidance to the operators to bring the process back to the optimal performance.In this paper,the cyanide leaching and zinc powder replacement process are taken as the research object.Comprehensive economic benefit is taken as the assessment index of whole process.The assessment model is established by using the Rough Set theory,and the effectiveness of the method is verified by simulation.